More than 30 million injuries are treated annually in hospital emergency departments (ED), and injuries account for more than 170,000 deaths in the U.S. In a well-established regional trauma care system, injured patients are expected to be transported to the most appropriate destination hospital ED. This ideal situation does not always happen in the U.S. so many patients with serious injuries are not triaged appropriately in the field. Furthermore, a high proportion of severely injured patients who are treated in EDs of non-trauma centers are not triaged to a higher level trauma center (undertriage). While there is strong evidence that trauma patients treated at trauma centers have a significantly lower mortality than if treated at non-trauma centers, little is known about the effect of undertriage on outcomes of trauma patients who were treated at level III or non-trauma centers. One big challenge in comparative trauma outcome study is the imbalance of risk factors between comparison groups because trauma patients are not randomly assigned to hospital EDs. Propensity scores have been used in recent years to adjust for such imbalance in observational trauma care studies. Currently, the methodology of using propensity score in comparing two treatment groups is well established. However, the methods for comparing across multiple groups are less well developed. This R03 research study will enhance the propensity score methods for multiple group comparisons and will apply the enhanced propensity score methodology in a national study of undertriage of trauma patients at level III and non-trauma hospitals. Two specific aims are: Aim 1) to develop and test propensity score methods for comparing ED and inpatient mortality of trauma patients treated at level I, II, III trauma centers and non-trauma centers (multiple group comparisons); and Aim 2) to study characteristics of undertriage and the effect of undertriage on mortality at level III and non-trauma centers. The largest all- payer ED database that is publicly available in the U.S., the Nationwide Emergency Department Sample (NEDS), will be used. This study proposes an innovative approach in assessing trauma care outcomes at EDs of different level trauma centers in the U.S. and will investigate characteristics of undertriage and its impact on mortality in trauma patients at the national level. Results will fill important knowledge gaps in characteristics and the mortality ris of undertriaged trauma patients in the U.S. This study addresses the mission of comparative effectiveness portfolio of the Agency for Healthcare Research and Quality and will provide evidence-based information to patients, clinicians, and other decisions-makers about the mortality risk of undertriaged trauma patients.